Hi Craig,
Comments inline.
Craig Feuerherdt wrote:
So the original questions remain:
Are the issues associated with having multiple polygon geometries in
the one table (unlikely in my opinion)?
Nope. Although perhaps to narrow the scope for testing, try putting a
filter on your evc table perform your query 3 times. Once with a single
polygon that you know that lies completely within your buffered
polygon, once with one that lies on the boundary, and once with one
that lies completely outside. Make sure you get the clipped results
your are expecting to see each time. This way you're only testing two
polygons (one from evc and your buffered polygon) and not your whole
dataset.
Is it something to do with the query?
Your query looks fine. I assume that your filter of buf_dist=1 returns
a single polygon? But try the filter I suggested, thus breaking your
problem into small pieces.
Are these issues associated with PostGIS or GEOS?
Unlikely. Having said that, once in a while we do stumble upon a weird
case that reveals a small bug in GEOS. More often than not, however,
the problem lies in the input geometries not being valid.
Is there a work around to solve the issue?
I noticed in your pictures two small slivers in your evc layer (one
around the middle left, and one around the middle top). Are you sure
your geometries are valid? I don't know if this is how your data is
supposed to look like, or if some geometries may be invalid, or if its
the way QGIS is rendering your data. I haven't used QGIS before, but I
know that JUMP does not render self-intersecting polygons very well
(sometimes it doesn't even try). Maybe something similar is happening
here?
Thanks in advance,
Craig
I think this may be an input geometry issue. I would verify that your
polygons in your evc and bnd_buff are valid and don't have duplicate
coincident points. Try running them through snaptogrid(the_geom,
<precision>).
Also, can you verify that your clipped resultset are all polygons?
Maybe QGIS has problems rendering a geometry collection?
Hope this helps,
Kevin
|
_______________________________________________
postgis-users mailing list
[email protected]
http://postgis.refractions.net/mailman/listinfo/postgis-users
